【作者】 孙秀云；

【导师】 方勇纯；

【作者基本信息】 南开大学 ， 控制理论与控制工程， 2013， 博士

【摘要】 小型无人直升机具有悬停、垂直起降与超低空飞行等灵活的飞行方式,因此它在军事与民用领域具有广泛的应用前景,并逐渐成为一个研究热点。然而,小型直升机系统的复杂性给其完全自主飞行带来很大的困难,同时也阻碍了它的快速发展,尤其我国的小型直升机技术研究相对于国外发达国家来说起步较晚,所以近些年来,我国科研工作者在改善小型直升机的自主飞行性能方面进行了深入探索。考虑到控制策略对自主飞行性能的重要影响,本文针对小型直升机系统中的强非线性、强耦合性、模型不确定性以及干扰等问题,提出了几种高性能的自主飞行控制算法。具体而言,本文的主要工作包括以下四个方面：首先,设计了一种具有指数收敛性能的小型无人直升机分层控制器。传统的分层飞行控制器通常忽略了姿态响应过程,假设姿态角能够迅速到达平移控制器输出的姿态给定值,这必定对控制质量产生影响。针对该问题,本文分析了姿态响应的耦合过程,采用四元数描述了小型直升机的姿态误差,并基于这种描述方法设计了内环的姿态非线性控制器,同时也避免了姿态运动的奇异性。平移运动控制器采用反步法进行了设计。通过李雅普诺夫方法证明了该分层控制器的指数收敛性能。仿真结果也表明该分层控制器具有良好的控制性能。第二,在考虑了主旋翼和副翼挥舞动态模型的基础上,应用反步法设计了自适应姿态与高度控制器。在实际工作时,由于不同飞行任务携带的不同负载会改变系统参数,因而需要重新调整控制器的增益,这非常不利于实际应用。为此,本文通过对小型无人直升机姿态动力学模型进行等效变换,使模型中的未知参数满足线性参数化条件,同时充分考虑主旋翼和副翼挥舞动作的影响,采用自适应策略设计了飞行控制器,较好地解决了不同工作条件下,小型无人直升机系统的质量和惯性矩阵等参数的不确定性问题。应用李雅普诺夫稳定性理论和芭芭拉引理证明了闭环控制系统的误差渐近收敛到原点处。仿真结果也表明该控制器在参数变化时依然能够保持良好的控制效果。第三,针对测量环节中的持续干扰,设计了具有时变柔化因子的模型预测姿态通道控制算法。这种控制方法将小型直升机姿态通道的传递函数模型进行等效变换,从而得到嵌入了积分环节的状态空间模型,并以此来减小闭环系统的稳态误差。接着综合分析了姿态通道系统的初始响应速度和调节时间两个方面,设计了时变柔化因子方法,用来抑制测量环节中持续干扰对柔化轨迹的影响。基于上述内容,考虑了舵机的输出限制,在控制输入无约束和有约束两种情况下,分别设计了模型预测姿态通道控制器,并利用线性系统理论分析了闭环系统的稳定性。将所设计的控制器应用于小型直升机实验平台,实验结果表明该控制器可以加快系统响应速度,减小模型不确定性与持续干扰等因素对系统性能的影响。第四,论文提出了一种基于非最小状态空间的姿态通道模型预测控制器。考虑到测量环节中存在持续干扰时,状态观测器性能降低的问题,本文建立了姿态通道系统的非最小状态空间模型,该模型的状态量由姿态通道的控制输入量与输出量构成。在此基础上,针对控制输入无约束和有约束两种情况,分别基于非最小状态空间模型设计了姿态通道模型预测控制器。利用线性系统理论证明了闭环系统在跟踪控制与抑制干扰方面具有良好的性能。飞行实验表明该控制算法具有良好的镇定与跟踪控制效果。更多还原

【Abstract】 A small-scale unmanned helicopter has many specific flight modes, such as hovering, vertical take-off and landing, flying at a low altitude, and so on. Owing to its flexibility, helicopters are widely utilized in both military and civil fields. However, some difficulties in autonomous flight, which are caused by the complexity of helicopters, limit its rapid development. Especially, in contrast with the developed countries, studies on the unmanned helicopter starts late in China. Therefore, Chinese scientists have devoted much effort to enhance its autonomous flight performance. Considering the important effect of control strategies on flight performance, the dissertation proposes several high performance control algorithms for the nonlinearity, coupling, model uncertainty characteristics and disturbance of the small-scale unmanned helicopter system. Specifically, the major contribution of this work can be summarized as follows.Firstly, a hierarchical flight controller with exponential convergent performance is designed for small-scale unmanned helicopters to deal with the coupling of the attitude dynamics and translation dynamics. The traditionally hierarchical flight controller generally ignores the attitude response process, and assumes that the attitude quickly reaches the set point developed by the translation controller, which usually reduces the control performance. In view of the above problem, the dissertation analyzes the couple process of attitude response, and describes the attitude error by the quaternion to avoid singularity. Based on the description, an attitude nonlinear controller is proposed. In addition, the translation controller is designed by utilizing the backstepping method. The Lyapunov method is employed to prove the stability of the proposed approach. Numerical simulations are included to verify its control performance.Secondly, considering the main and aileron rotor flapping dynamics, an adaptive attitude and altitude controller is designed using the backstepping method. To achieve a satisfactory flight performance, it is important to tune suitable control gains for different flight missions in which the helicopter carries different loads. However, the tuning process is usually complicated for helicopter users. Therefore, we firstly make the unknown parameters meet the linear parametric conditions by the equivalent transformation of the attitude dynamics model. Subsequently, by analyzing the rotor flapping dynamics, an adaptive control algorithm is designed for the mass and inertia matrix unknown parameters from different loads. Lyapunov theory and Brabalat’s lemma are further utilized to prove that all the closed-loop system errors converge to zero asymptotically. Simulation results show that the designed controller still maintains good control results in the presence of the system parameters variation.Thirdly, a discrete-time model predictive attitude channel controller with the time-varying softening factor is proposed for the persistent disturbances that enter the system from the measurement components. In order to reduce the static error, we transform attitude channel transfer functions into the state space functions imbedded the integral. Then we carefully analyze the response speed and adjusting time of the attitude channels, and propose the time-varying softening factor to suppress the persistent disturbances effect on the reference trajectory. Moreover, considering the limited operation range of the helicopter rudder, the discrete-time model predictive attitude channel controllers are designed for both limited and unlimited control inputs. The stability of the close-loop system is proven by considering the fact that all the eigenvalues lie inside the unit circle. The designed controllers are then employed in a small-scale helicopter system, and the results show that it enhances the speed of the control system response, and decreases the influence of model uncertainty and persistent disturbances remarkably.Finally, the dissertation proposes a discrete-time model predictive attitude channel controller based on the non-minimal phase state space model. In the model predictive control system with the persistent measurement disturbances, the performance of state observer is reduced. Owing to this problem, we develop the non-minimal phase state space model for the attitude channels of a small-scale helicopter, whose states consist of the attitude channel control input and output data. Then based on the non-minimal phase state space attitude channels model, the discrete-time model predictive attitude channel controllers are designed for both constrained and unconstrained control inputs. The linear system theory proves the tracking performance and anti-jamming ability of the controllers. Practical flight experiments illustrate the satisfactory performance of the proposed controllers.更多还原